- Email: [email protected]
Chemical characterisation of rainwater at Stromboli Island (Italy): The effect of post-depositional processes
Accepted Manuscript Chemical characterisation of rainwater at Stromboli Island (Italy): The effect of post-depositional processes
Marianna Cangemi, Paolo Madonia, Rocco Favara PII: DOI: Reference:
S0377-0273(16)30448-6 doi: 10.1016/j.jvolgeores.2017.01.023 VOLGEO 6009
To appear in:
Journal of Volcanology and Geothermal Research
Received date: Revised date: Accepted date:
14 November 2016 27 January 2017 28 January 2017
Please cite this article as: Marianna Cangemi, Paolo Madonia, Rocco Favara , Chemical characterisation of rainwater at Stromboli Island (Italy): The effect of post-depositional processes. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Volgeo(2017), doi: 10.1016/j.jvolgeores.2017.01.023
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Chemical characterisation of rainwater at Stromboli Island (Italy): the effect of post-depositional processes Marianna Cangemia, Paolo Madoniaa*, Rocco Favaraa Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Palermo, via Ugo La Malfa 153, 90146 Palermo (Italy) *
Corresponding author, Email: [email protected]
PT
Abstract
Volcanoes emit fluids and solid particles into the atmosphere that modify the chemical composition
RI
of natural precipitation. We have investigated the geochemistry of Stromboli’s rainfall during the
SC
period from November 2014 to March 2016 using a network of a new type of sampler specifically designed for operations on volcanic islands. We found that most of the chemical modifications are due to processes occurring after the storage of rainwater in the sampling bottles. These processes
NU
include dissolution of volcanogenic soluble salts encrusting volcanic ash and a variable contribution of sea spray aerosol. Our data showed noticeably less scatter than has previously been achieved with
MA
a different sampling system that was more open to the atmosphere. This demonstrates the improved efficacy of the new sampler design. The data showed that post-depositional chemical alteration of rain samples dominates over processes occurring during droplet formation ad precipitation. This has
PT E
D
important implications for the calculation of fluxes of chemicals from rainfall in volcanic regions.
Keywords
1.
CE
Rainwater chemistry; sea spray; plume; volcanic ash; Stromboli; post-depositional processes.
Introduction
Volcanoes are important sources of gases, solid particles and aqueous acid droplets emitted at high
AC
temperature into the atmosphere (Cadle, 1980; Oppenheimer, 2003; Mather et al., 2003; Gerlach, 2004; Textor et al., 2004; Von Glasow et al., 2009). Atmospheric precipitations represent a significant mechanism for maintaining the balance between sources and sinks of particulate, providing a natural scavenging of the chemical dispersed into the atmosphere also in terms of acidification of water (Gonzáles and Aristizábal, 2012). The interaction between volcanic activity and rain chemistry has been investigated worldwide, and among the others at Kilauea, Hawai (Siegel et al. 1990; Scholl and Ingebritsen 1995), Soufriere Hills volcano, Monserrat (Edmonds et al. 2003), and in Italy at Mt Etna (Bellomo et al. 2003), Vulcano island (Liotta et al. 2006; Madonia and Liotta, 2010), Stromboli island (Bellomo et al., 2003; Liotta et al., 2006; Madonia et al., 2013), and Mt Vesuvius (Madonia and Liotta,
1
ACCEPTED MANUSCRIPT 2010). Rain droplets interact with volcanic products giving rise to rainwater with a strong chemical fingerprint of the dissolution of soluble ions and gases (Bellomo et al., 2003; Liotta et al. 2006; Madonia and Liotta, 2010). In addition, volcanic aerosol can act as cloud condensation nuclei promoting cloud formation. The deposition of precipitations chemically altered by the interaction with volcanogenic volatiles and particles impacts the surrounding environment, damaging vegetation (Delmelle et al. 2002), changing the soil chemistry (Delmelle et al. 2003) or affecting water supply systems (Cuoco et al., 2013a, b;
PT
Madonia et al., 2013).
During the last decades several studies focused on the chemistry of rainfalls in volcanic areas have
RI
been carried out on Italian active and quiescent volcanoes, because these are inserted into heavy
SC
urbanized contexts and the potential impacts on human health are highly significant (Bellomo et al., 2003; Liotta et al., 2006; Madonia and Liotta 2010; Madonia et al., 2013). Among these areas, Stromboli is particularly relevant because of both its continuous volcanic activity and the presence of
NU
a very populous human community, especially during the touristic season. A previous work on the geochemistry of Stromboli’s rainwater was conducted by Liotta et al. (2006).
MA
The authors investigated the chemical and isotopic compositions of the precipitation between October 2003 and October 2005, finding no direct contribution of volcanogenic H or O. The chemical composition of rainwater close to the coastline was found influenced by the sea aerosol, whereas the
D
chemical effects due to the volcanic activity were found in those samples collected closer the summit
PT E
vents. Moreover, the S/Cl, Cl/F, and S/F molar ratios were consistent with those measured in the volcanic plume. Minor and trace element concentrations in rainwater were not considered during this study and, more in general, in other researches in Mediterranean volcanic islands. The reason is
CE
that the protocols for collecting rainwater samples for minor and trace element determination are based on Büchner type funnels with a single sampling session covering a time span of about 15 days (Calabrese et al., 2011). This procedure does not fit the economical and logistic requirements for
AC
studies in insular areas, like Stromboli, where it is really difficult ensuring a so frequent and constant field activity.
Following these premises we designed a new type of rain sampler for the determination of minor and trace elements, able to properly collect samples over several months, and tested it at Stromboli island. We choose this location because previous studies characterised the chemical composition of volcanic ash (Bagnato et al., 2011) and rainwater collected both directly from the atmosphere (major elements only; Liotta et al., 2006) and harvested by roof catchment systems (Madonia et al., 2013). Moreover, during our sampling campaign, from November 2014 to March 2016, Stromboli was experiencing a period of different volcanic activity with respect to the previous study by Liotta et al.
2
ACCEPTED MANUSCRIPT (2006), giving us the opportunity to look for its possible proxies in the geochemical characterisation of precipitation.
2. Volcanological setting Stromboli, the north-easternmost island of the Quaternary Aeolian volcanic arc off the northern coast of Sicily (Southern Italy), is a stratovolcano that rises from a depth of ~2000 m b.s.l. in the Tyrrhenian sea with a subaerial portion 924 m high (Fig. 1).
PT
It is characterised by the so called “Strombolian activity”, consisting of passive magma degassing alternated to short (a few to few tens of s) 100- to 200-m high scoria-rich jets produced by explosions
RI
of variable energy every 10–20 min (Patrick et al., 2007; Andronico et al., 2008; Taddeucci et al.,
SC
2012). This activity take place from vents lying on the NE-SW-trending crater terrace, located at about 750 m above sea level (a.s.l.) in the upper portion of the Sciara del Fuoco. The volcanic plume continuously emitted by the vents is mainly composed of water vapour (over 50
NU
wt%), with CO2 and SO2 as other principal gaseous species; HCl and HF are also emitted as secondary
3.
MA
components (Allard et al., 1994).
Sampling and analytical methods 3.1
Sampling methods
D
As previously introduced we realised a new sampler for collecting rainwater samples for trace
PT E
element analysis, designed for properly working over several months: the main issues in using the normal Büchner type funnels (Calabrese et al., 2011) are the reduced volume of the sampling bottle and its insufficient isolation from the atmosphere. For solving these problems, we designed and
CE
realized a rain sampler where the funnel diameter/sampling bottle volume ratio allows collecting up to 520 mm of rain; based on literature data (Cicala, 2000), this height corresponds to the total precipitations falling at Stromboli during several months of the autumn-winter wet season.
AC
Considered the extended duration of the sampling period, we adopted an array of 3 sequential filters (Fig. 2) instead of the single filter usually equipping the Büchner type funnels: in this way we reduced possible evaporation losses and contamination due to the input of fine particulate into the sampling bottle. About the construction details, our system consists of a 2 l Nalgene bottle over which a composite TeflonTM funnel, with a diameter of 75 mm, is fasten. The funnel can be separated in each component for easy and total cleaning. A sequence of three pierced disks with a downward decreasing hole diameter (5, 4, 3 mm respectively) is inserted inside the funnel. Sheets of filter paper (Whatman® Cellulose quantitative filter paper, ashless, grade 40) are put between each pierced disks
3
ACCEPTED MANUSCRIPT with the aim of avoiding the clog of the holes, preventing ash entering the sampling bottle and consequently minimizing the prolonged interaction between volcanic ash and rainwater after collection. Such a system is also useful for reducing evaporation of water after the collection: in fact, in normal Büchner funnel there is only one filter that is strongly heated by the incident solar radiation. In our system the multiple disks work as a radiative shield, limiting heat transfer to the interior of the bottle and evaporation of the collected water. Moreover, the rain sampling system is put into a plastic case, with an air gap between the case and the bottle for further reducing radiative
PT
heating. The whole system is mounted on a metallic pole at about 1.5 m above the ground, ensuring the horizontality of the catchment area of the funnel. All the equipments were prewashed with a 2%
RI
nitric acid (Merck Suprapur) solution, and then rinsed several times with de-ionized (MilliQ) water,
SC
dried under hood, packed in double clean plastic bags, and zipped until exposure in the field. At the end of the sampling period, the sampling bottle was substituted with a new clean one, capped and transported in the laboratory for analysis.
NU
A total of 3 rain sampling systems were installed at different distance from the crater and with different directions respect to the dominant winds (Fig. 1; Table 1). The first prototype was installed
MA
at the COA volcanological observatory in November 2014 and, after a preliminary test on its performances, we built up the sampling network with another system installed at FRT (in the same location of Liotta et al., 2006) in January 2015. We completed the network in August 2015 with a
D
third system in the Piscità area (DER, Fig. 1). The sampling sites are distributed both at variable
PT E
distances from the summit vents and at variable angular distances with respect to the main volcanic plume propagation direction, allowing us to estimate the possible interactions between volcanic activity and rainfalls. Moreover, 2 single rain events were collected at SCA (close to Stromboli’s
respectively.
CE
harbour) and PDR (few hundreds of metres northernmost of COA) in March and May 2015,
3.2 Analytical methods
AC
Chemical-physical parameters (pH and electric conductivity) were measured directly in the field using a Thermo Orion instrument and a 8102BN Ross combination pH electrode. Water samples used for the determination of dissolved major and trace elements were first filtered using 0.45 µm Millipore MF filter and then: a) collected in LD-PE (low-density polyethylene) bottles for major element analyses, acidifying to ca pH 2 the aliquot destined to cation determination with HCl; b) collected in PP (polypropylene) bottles for trace element analyses, acidified to pH ca 2 with ultrapure concentrated HNO3. Untreated aliquots were stored for alkalinity determinations, made via titration with HCl (0.1 N). Electric conductivity and pH were measured in the field using Orion instruments equipped with Hamilton electrode (pH). Major ions were determined by ionic chromatography using
4
ACCEPTED MANUSCRIPT Dionex columns AS14 and CS12 for anions and cations respectively. Major element data are presented in Table 2. The trace elements, were analysed by ICP-MS (Agilent 7500ce) equipped with a Micromist nebulizer, a Scott double pass spray chamber, a three-channel peristaltic pump, an auto sampler (ASX-500, Cetac), and an octopole reaction system (ORS) to remove the interferences. The mass spectrometer was calibrated with a multi-element standard solution with 11 calibration points. A multi-element standard solution was prepared by dilution from single and multi-element stock solution,
PT
respectively at 1.000 mg/L and 100 mg/L. Multi-element stock calibration solutions were always prepared freshly in 10mL polyethylene tubes by dilution using 1% nitric acid for trace analysis as
RI
diluent.
SC
The sensitivity variations were monitored using 103Rh, 115In, and 185Re, internal standards were added directly online by an appropriate device that mixes the internal standard solution into the sample just before the nebulizer. The concentration of the internal standard was about 10 μg/L in all sample
NU
and calibration solutions. The precision of the analysis was checked by running five replicates of every standard, and the sample was always within ±10%. Data accuracy was evaluated by analysing
MA
standard reference materials (Spectrapure Standards SW1 and 2, SLRS4, NIST 1643e, Environment Canada TM 24.3, and TM 61.2) for each analytical session, and the error for each element was <15%.
Results
Chemico-physical parameters and concentration of major elements
PT E
4.1
D
4.
In order to evaluate the spatial variations of pH and TDS we plotted these values against three different parameters expressing the possible effect of volcanic activity, and in particular of the
CE
interaction with the volcanic plume. These are the planar, vertical and angular distances of the sampling sites with respect to the summit vents, expressed in terms of relative distances, altitude of the sites, and angles formed between the propagation direction of the plume and the line connecting
AC
the vents with the sampling point. Although only 3 points are available, thus not giving the opportunity of calculating statistically significant relationships, useful qualitative information can be retrieved. These relationships are illustrated in Fig. 3a-I, showing the variations of pH (a-c) and TDS (d-f) with distance from the summit vents, altitude and angle with respect to plume direction. The pH of rainwater was highly variable, ranging from 3.0 to 5.3 (Table 2); as shown in Fig. 3a-c, correlations are found with altitude and angle, while the relationship with distance is not so clear. The closest site to the summit vents (FRT), at highest altitude and the lowest angular distance from the propagation direction of the plume, was characterised by the most acid water (pH range 3.0-4.2). On the contrary, the highest pH values were recorded at DER (range 4.1-5.3), that is the lowest altitude and at the
5
ACCEPTED MANUSCRIPT maximum angular distance from the propagation direction of the plume. The COA site shows intermediate values. Similar relationships are found for total dissolved solids (TDS), with highest values at FRT and lowest at DER with COA in an intermediate position (Fig. 3d-f). About the temporal variations of these parameters, it is difficult discriminating between their possible dependence on duration of sampling intervals or total amounts of precipitation, because these two are also strongly correlated (r2= 0.97). Anyway, Fig. 4a highlights a progressive decrease of
PT
pH following the increase of the sampling period (r= -0.65); two single rain events (PDR and SCA) are also reported for comparison. Similarly, TDS negatively correlates with rainfall amount (Fig. 4b).
RI
Major element concentrations are listed in Table 2; electroneutrality balance is not assured because
SC
not all the ionic species were determined but only those of interest for the aim of the present study. In the modified Langelier-Ludwig diagram, shown in Fig. 5a, the sampled rainwaters are aligned along a compositional trend produced by different mineralization processes. The two end members are sea
NU
spray aerosol and a sulphate-dominated term, typical of the Stromboli volcanic environment (Liotta et al., 2006). Groundwater sampled into wells located in the coastal area of the island (Grassa et al.,
MA
2008) lies along this curve, close to the seawater composition. A difference is observed among our and previous (Liotta et al., 2006) data. The former show a good alignment, whereas the latter are much more scattered, especially for compositions close to the volcanic end-member. Some intra-
D
annual compositional fluctuations are also observed and shown in Fig. 5b, related to the COA site for
PT E
which the longest time series, covering more than a complete year, is available. We plotted Cl/SO4 ratio and F concentrations vs time: the two variables are anti-correlated and indicate an increase in fluoride concentration during the dry season, coupled to a relative enrichment of SO4 with respect to
CE
Cl (see Table 2). Fluoride also varies in space, as shown in Fig. 3g-I, as it follows the same pattern of TDS: the smallest are the planar (with the exception of COA) and vertical distances from the summit vents and the angle formed between the propagation direction of the plume and the line connecting
AC
the vents with the sampling points, the highest is F concentration. It is relevant that F concentration often exceeds the safety limits for drinking water, fixed at 1.5 mg l-1 by the World Health Organization (WHO, 2011), at FRT and COA, with a maximum value of 11.9 mg l-1 measured in the August 2015 sample at FRT. The scatterplots of Fig. 6 (a-d) compare pairs of concentrations of the most representative ions in rainwater (Na, K, Ca, Cl, SO4) with respect to those of volcanic plume (Allard et al., 2000), Mediterranean seawater and the mineralogical phases by whose dissolution these can derive. As a general remark, all the rainwater samples are comprised in an area delimited by the lines representative of the possible chemical end-members.
6
ACCEPTED MANUSCRIPT 4.2
Minor and trace metal concentrations
Metal concentrations are listed in Table 3 and illustrated in Fig. 7 in roughly order of relative abundances. These range from about 0.005 g l-1 for ultra-trace elements (e.g. Be, Cs, Mo, U) up to ≥1000 g l-1 for minor elements (Al, Li, Zn). The concentration patterns are similar to those found by Madonia et al. (2013) in rainwater collected in roof catchment systems and Bagnato et al. (2011) in ash leachate, with intermediate absolute values between these end members. The safety limits fixed by the WHO for drinking water are also reported in Fig. 7. As shown in the
PT
diagrams, all the concentrations of the determined elements fall below these limits, with the exception of Sb that exceeds its threshold in all the sites.
RI
As expected for metals in aqueous solution, the concentration of quite all the determined elements
SC
negatively correlate with pH (Table 4). Exceptions are represented by Zn and Mo. For better constraining the relative amount of each element in rainwater with respect to its concentration in volcanic ash, we calculated the enrichment factor (EF) for each X element, following
NU
the relation (Zoller et al., 1983):
EFash = (X/Ti)rain/(X/Ti)ash; EFash leachate = (X/Ti)rain/(X/Ti)ash leachate
MA
We used Ti as reference element due its low volatility under magmatic conditions and low mobility during weathering processes (Aiuppa et al., 2000). EFs are shown in Figure 8 ordered according to their mobility. Sampling sites show enrichments in Mo, Zn, Tl, As, Cu and Ni, with respect to ash
D
(Allard et al., 2000; Bertagnini et al., 2003). Elements like Ba, Sr, Rb, Al, Fe, Li, and Cr have EFs values
5.
Discussion
PT E
below the unity.
CE
Due to the logistic and economic reasons discussed in the introduction, data about minor and trace elements in rainwater from volcanic islands are rare worldwide and absent from the Mediterranean.
AC
Ours are the first available data from a Mediterranean volcanic island, thanks to a sampling system specifically designed for collecting multi-months samples. A first controversial question, already arisen from previous studies devoted to the geochemical characterization of rainwater in volcanic areas, is whether the chemical composition of the sampled precipitation is driven by interaction processes occurring in the atmosphere, or the sample is chemically modified after the collection in the sampling bottle. The question is not captious, because if the second hypothesis is true, no quantitative inferences about the flux of volcanogenic elements in the atmosphere can be extracted using as a proxy the chemical composition of rainwater. In a study based at Vulcano Island, Madonia and Liotta (2010) demonstrated that the interaction with acidic gases, emitted by a fumarolic field, is able to strongly modifying the chemical fingerprint of
7
ACCEPTED MANUSCRIPT rain samples after collection. These authors exposed to the atmosphere a bottle containing MilliQ water in a location few hundreds of meters downwind of the fumarolic field of La Fossa cone, finding over 20 milliequivalents of dissolved sulphate after 2 months of exposition. A similar behaviour is found in the chemistry of the samples collected during our study. As shown in Fig. 4a, the pH of the collected rainwater negatively correlates with the duration of the sampling period, suggesting that the water collected in the bottle progressively reacts with volcanic acidic gases (Delmelle, 2003) continuously emitted from the summit vents. In fact, as illustrated in Fig. 3a-c,
PT
the closest a sampler is located to areas affected by interactions with volcanic plume (FRT) the highest is the acidity of the collected rainwater. This must not be intended as a planar distance,
RI
because the graphs indicate that the direction of the plume (Fig. 3c) and the elevation proximity (Fig.
SC
3b) are more effective in controlling pH. In the same way, an opposite correlation is found for TDS (Fig. 3d-f) because the most acid is the water the higher is its capability of scavenging chemicals from ash particles. The fingerprint of this process is found in the exponential decrease of TDS with the
NU
increase of rain amount (Fig. 4b). Chemical composition of rainwater is controlled by the leaching of soluble salts, both coated on the surface of volcanic particles accumulated into the sampler funnel
MA
(outside the multiple filter system) or coming from the deposition of sea spray. When the volume of water accumulated into the bottle is low, the dissolution of these particles leads to a higher content of dissolved ions. Conversely, when the water volume is higher, the final ion concentration will be
D
lower. These observations are confirmed by the minor and trace element behaviour. As shown in
PT E
Table 4, the concentrations of the various species negatively correlate with both amount of rainfall and duration of sampling periods.
The novel inference from our study is that the chemistry of rainwater sampled in active volcanic
CE
areas with wet and dry rain gauges is primarily influenced by reactions occurring after the collection of rainwater into the samplers. We were able to obtain this data because we collected for the first time rainwater samples representative of the total amount of precipitation in several months. Lot of
AC
caution should then be adopted for extracting quantitative inferences on the atmospheric emission of volcanogenic elements based on their concentration in rainwater samples. Although quantitative flux models seem difficult to be developed, interesting qualitative information concerning the origin of the elements and their variation with time can be extracted from the chemical composition of rainwater collected at Stromboli. A double origin for the dissolved ions, sea spray and volcanogenic compounds, is supported by the Langelier-Ludwig diagram (Fig. 5a). It is noteworthy that SO4 is enriched with respect to Cl during the spring-summer season, as well as fluorine (Fig. 5b), indicating a stronger volcanic fingerprint during these months. This effect could be driven by the coaction of two different processes: i) in winter
8
ACCEPTED MANUSCRIPT winds are stronger and sea waves higher, producing more sea spray that is transported and deposited into the funnels with respect to the spring-summer season; ii) during winter atmospheric relative humidity is higher than in warm season. Since acidic species are strongly hydrophilic these are quickly removed from the plume, due to the interaction with the atmospheric water droplets. As a consequence, the residual volcanic plume interacting with rainwater accumulated into the samplers will be depleted in acidic species. This last consideration opens a question about a possible implication for remote sensing of volcanic plume composition, since this scavenging may also be
PT
active during these measurements.
The dissolution of the volcanogenic coating material encrusted on volcanic ash is the other major
RI
source for chemicals found in rainwater. This material is mainly composed of CaSO4 (anhydrite or
SC
gypsum), halite and sylvite, formed during the reactions between acidic gases (mainly SO2g and HClg), abundant in the plume, and the volcanic particles (Rose, 1977; Oskarsson, 1980; Hinkley and Smith, 1982; Delmelle et al., 2007; Bagnato et al., 2011; Barone et al., 2016). These minerals have also been
NU
found in volcanic ash from Mt Etna (Barone et al., 2016). These dissolution processes are clearly identified in the binary diagrams of Fig. 6. In the SO4 vs Cl plot (Fig. 6a) rainwater is comprised in a
MA
band delimited by the lines representative of plume and seawater compositions. The Ca vs SO4 (Fig. 6b), Na vs Cl (Fig. 6c), and K vs Cl (Fig. 6d) diagrams evidence that the dissolution of anhydrite, halite and sylvite and the interactions with sea spray and plume explain the ratios found among these
D
compounds in the analysed rain samples.
PT E
The comparison between our data and those collected by Liotta et al. (2006) gives insights about the efficiency of our sampling system. Since that study was aimed to retrieve chemical and isotopic information from the collected samples, different samplers were used. These were made of 25-l
CE
cylinder shaped PET tanks with a 30-cm-diameter PET funnel, filled with Vaseline oil (250 cc) in order to avoid evaporation of the collected rain (more details about this system can be found in the paper by Madonia and Liotta, 2010). It is noteworthy that such a system is not equipped with any filter able
AC
to avoid the input of solid particulate inside the collection tank. A different behaviour is shown by our data in the Langelier-Ludwig diagram (Fig. 5) with respect to those from the previous study by Liotta et al. (2006), with the former better aligned following the trend having seawater and volcanic end members with respect to the latter. This difference suggests that our sampling system works better, because the multi-paper filters mounted on the funnel avoid the entrapment of solid volcanic particulate into the sampling bottles. The leaching of this particulate by the rain collected into the samplers could be responsible of the major variability of the chemical composition found by Liotta et al. (2006). Truthfully, a different explanation could be attributed to a more variable volcanogenic source during the years 2003-2005 (data from Liotta et al., 2006) with respect to a most recent
9
ACCEPTED MANUSCRIPT period (2014-2016, our data). Both sampling periods followed eruptive phases, occurred at Stromboli between December 2002 and July 2003 and August and November 2014, respectively. The two eruptions presented different styles: the 2002-03 event was energetic and showed an evident overpressurization of the shallow fracture system that led to a paroxysmal explosion on 5 April 2003, while the most recent event was more “quiet” due to a lesser amount of volatiles involved (Rizzo et al., 2015, and references therein). This difference in the degassing style seems being reflected in the two datasets. A more turbulent regime characterised the period 2003-05, leading to variable ratios
PT
between volcanogenic and sea spray elements. By contrast, a more invariant volcanogenic source dominated the period 2014-16, resulting in a quite constant chemical composition of rainwater.
RI
A special attention deserves F concentration. This element often exceeds the safety limit established
SC
by the WHO for drinking water in COA and FRT sites (Table 2), suggesting that the collection and storage of rainfalls for sanitary and alimentary purposes could imply human health issues. Nevertheless, the previous study by Madonia et al. (2013) on the quality of water collected in roof
NU
water catchment systems (RWCS) at Stromboli indicated that the RWCS water is generally close to the optimal concentration of 1 mg/l. This discrepancy can be attributed to the post-depositional
MA
reactions between the volcanic plume and the rainwater collected in the samplers; the higher F concentrations found in the present study are reasonably due to the small volume of water contained in the sampling bottles, strongly affected by the interaction with the plume (the highest F
D
concentrations are found in FRT site, which is the closest to the summit vents). Conversely, the water
PT E
stored in closed cisterns far away from the crater terrace is less contaminated by volcanogenic fluorine.
About trace elements, a first evidence is a lower content of Sr and Rb in rainwater with respect to
CE
samples collected in roof catchment systems (Madonia et al., 2013): this is likely due to the interaction between the harvested rainwater and the surface of the cisterns where these are stored, treated with lime. More generally, since the trace element pattern is similar to that found in ash
AC
leachate (Bagnato et al., 2011), leached volcanic ash is the most probable source for the elements found in rainwater.
Calculations of enrichment factors (Fig. 8) show that rainwater are enriched in those elements typical of ash that had interacted with the plume, like Mo, Zn, Tl, Cu, As, and Ni. These elements, present on ash surface under the form of halides and S compounds (e.g. CuSg, ZnCl2g, Krauskopf, 1964; Naughton et al., 1974; Moune et al., 2006; Varekamp et al., 1986), are easily leached by the rain. On the contrary, EFs below the unity characterize the so called refractory elements, less mobile during chemical weathering processes.
10
ACCEPTED MANUSCRIPT A final consideration concerns the possible issue for human health related to high concentrations of potentially toxic elements present in rainwater. As shown in Fig. 7 the only element exceeding the safety limit established by the WHO is Sb in all the sampling sites. However, as already discussed for F, this apparent high concentration could be caused by the post-depositional interaction with the volcanic plume in the rain samplers, which is not effective for water collected in RWCS; the comparison with the data by Madonia et al. (2013) confirms that rainwater harvested in cisterns is
6.
PT
always below the safety limits.
Conclusions
RI
First of all, our findings are relevant from a methodological point of view. Since the most of the
SC
water-rock interaction processes occur between rainwater and volcanic particles accumulated into the funnel, the use of wet and dry sampler should be mandatory in studies devoted to the chemical characterisation of natural precipitation in volcanic areas. Sampling systems based on the collection
NU
of bulk samples seem to be not adequate for such kind of researches, because it is impossible to discriminate between post-depositional scavenging of ash (leaching of solid particles accumulated
MA
into the funnels) and water-rock interaction processes occurring during rain droplets formation and precipitation (dissolution of gaseous and solid particles from volcanic plumes into the atmosphere). Although the two processes can not be distinguished, chemistry of rain collected in bulk samplers
D
could give useful, even though qualitative, information on the composition of a volcanic plume. In
PT E
this sense a bulk rain sampler could work as a passive trap for volcanic acidic species contained in the plume, which are tracers of the volcanic activity state. The more a volcano is active the more its plume is enriched in acidic species. At the same time, the more a plume is acidic the more it will be
CE
able to modify the chemical composition of the collected rain.
Acknowledgments
AC
This work was partially funded and realised in the behalf of the FSE postgraduate specialisation course “Valutazione e monitoraggio dei rischi ambientali naturali” (Regione Siciliana, Asse IV, Capitale Umano). We wish to thank our colleagues of INGV Paolo Cosenza and Giuseppe Riccobono, who improved the original project of the sampling system with their suggestions and materially realized the devices. We are also grateful to two anonymous reviewers and the editor Prof. Heidy M. Mader who gave us useful suggestions for improving this paper.
References Aiuppa, A., Allard, P., D’Alessandro, W., Michel, A., Parello, F., Treuil, M., Valenza, M., 2000. Mobility
11
ACCEPTED MANUSCRIPT and fluxes of major, minor and trace metals during basalt weathering at Mt. Etna volcano (Sicily). Geochim. Cosmochim. Acta, 64, 1827–1841, doi:10.1016/S0016-7037(00)00345-8.
Allard, P., Aiuppa, A., Loyer, H., Carrot, F., Gaudry, A., Pinte, G., Michel, A., Dongarrà G., 2000. Acid gas and metal emission rates during longlived basalt degassing at Stromboli volcano. Geophys. Res. Lett. 27(8), 1207–1210, doi:10.1029/1999GL008413.
RI
degassing budget of Stromboli volcano, Nature, 368, 326330.
PT
Allard, P., Carbonelle, J., Metrich, M., Loyer, H., Zettwoog, P., 1994. Sulphur output and magma
SC
Andronico, D., Corsaro, R.A.. Cristaldi, A., Polacci, M., 2008. Characterizing high energy explosive eruptions at Stromboli volcano using multidisciplinary data: An example from the 9 January 2005
NU
explosion. J. Volcanol. Geotherm. Res. 176, 541–550.
Bagnato, E., Aiuppa, A., Andronico, D., Cristaldi, A., Liotta, M., Brusca, L., Miraglia, L., 2011. Leachate
MA
analyses of volcanic ashes from Stromboli volcano: A proxy for the volcanic gas plum composition? J. Geophys. Res. 116:D17204, doi:10.1029/2010JD015512,2011
D
Barone, G., Mazzoleni, P., Corsaro, R.A., Costagliola, P., Di Benedetto, F., Ciliberto, E., Gimeno, D.,
PT E
Bongiorno, C., Spinella, C., 2016. Nanoscale surface modification of Mt. Etna volcanic ashes. Geoch. Cosmochim.
CE
Bellomo, S., D'Alessandro, W., Longo, M., 2003. Volcanogenic fluorine in rainwater around active
AC
degassing volcanoes: Mt. Etna and Stromboli Island, Italy. Sci. Total Environ. 301, 175–185.
Bertagnini, A., Métrich, N., Landi, P., Rosi M., 2003. Stromboli volcano (Aeolian Archipelago, Italy): An open window on the deep feeding system of a steady state volcano, J. Geophys. Res. 108(B7), 2336, doi:10.1029/2002JB002146.
Cadle R.D., 1980. A comparison of volcanic with other fluxes of atmospheric trace gas constituents. Rev. Geophys. Space Phys. 18, 746–752.
Calabrese, S., Aiuppa, A., Allard, P., Bagnato, E., Brusca, L., D’Alessandro, W., Parello, F., 2011. Atmospheric sources and sinks of volcanogenic elements in a basaltic volcano (Etna, Italy). Geoch.
12
ACCEPTED MANUSCRIPT Cosmochim. Acta 75/23, 7401-7425, doi: 10.1016/j.gca.2011.09.040.
Cicala, A., 2000. Guida alla Meteorologia delle Isole Eolie. Aldo Natoli Editore, Lipari, 149pp.
Cuoco, E., Tedesco, D., Podera, R.J., Williams, J.C., De Francesco, S., Balagizi, C., Darrah, T.H., 2013a. Impact of volcanic plume emissions on rain water chemistry during the January 2010 Nyamuragira eruptive event: Implications for essential potable water resources. J. Hazard Mater.
PT
581
RI
Cuoco, E., Spagnuolo, A., Balagizi, C., De Francesco, S., Tassi, F., Vaselli, O., Tedesco, D., 2013b.
SC
Impact of volcanic emissions on rainwater chemistry: The case of Mt Nyiragongo in the Virunga volcanic region (DRC). J. Geochem.
79.
NU
Delmelle, P., Stix, J., Baxter, P.J., Garcia-Alvarez, J., Barquero, J., 2002. Atmospheric dispersion, environmental effects and potential health hazard associated with the low-altitude gas plume of
MA
Masaya volcano, Nicaragua. Bull. Volcanol. 64, 423–434.
Delmelle, P., 2003. Environmental impacts of tropospheric volcanic gas plumes. In: Oppenheimer, C.,
PT E
D
Pyle, D.M., Barclay, J. (eds). Volcanic degassing, vol. 213. Geol. Soc. Spec. Publ. Lond. pp 381–399.
Delmelle, P., Lambert, M., Dufrene, Y., Gerin, P., Oskarsson, N., 2007. Gas/aerosol‐ash interaction in volcanic plumes: New insights from surface analyses of fine ash particles, Earth Planet. Sci. Lett. 259,
CE
159–170, doi:10.1016/j.epsl.2007.04.052.
Edmonds, M., Oppenheimer, C., Pyle, D.M., Herd, R.A., 2003. Rainwater and ash leachate analysis as
AC
proxies for plume chemistry at Soufriere Hills Volcano, Montserrat. In: Oppenheimer C, Pyle DM, Barclay J (eds). Volcanic degassing. Geological Society, London, pp 203–218
Gerlach, T.M., 2004. Volcanic sources of tropospheric ozone-depleting trace gases. Geochem, Geophys. Geosyst. 5 (Q09), 007.
Gonzáles, C.M., Aristizábal, B.H., 2012. Acid rain and particulate matter dynamics in a mid-sized Andean city: The effect of rain intensity on ion scavenging. Atmospheric Enviorn. 60, 164-171.
13
ACCEPTED MANUSCRIPT Grassa, F., Inguaggiato, S., Liotta, M., 2008. Fluid geochemistry of Stromboli. In: Calvari S, Inguaggiato S, Puglisi G, Ripepe M, Rosi M (eds) The Stromboli volcano: an integrated study of the 2002–2003 eruption. Am. Geophys. Union, Geophys. Mon.
Hinkley, T.K., Smith, K.S., 1982. Leachate chemistry of the tephra from the May 18 1980 eruption of Mount St. Helens, Eos Trans. AGU, 63(45), 1143.
PT
Krauskopf, K.B., 1964. The possible role of metal compounds in ore genesis, Econ. Geol., 59, 22–45,
RI
doi:10.2113/gsecongeo.59.1.22.
SC
Liotta, M., Brusca, L., Grassa, F., Inguaggiato, S., Longo, M., Madonia, P., 2006. Geochemistry of rainfall at Stromboli volcano (Aeolian Islands): isotopic composition and plume-rain interaction.
NU
Geochem. Geophys. Geosyst. 7Q07006. doi:10.1029/2006GC001288
Madonia, P., Liotta, M., 2010. Chemical composition of precipitation at Mt Vesuvius and Vulcano
MA
Island, Italy: volcanological and environmental implications. Environ. Earth Sci. 61, 159-171.
Madonia, P., Cangemi, M., Bellomo, S., D’Alessandro, W., 2013. Influence of volcanic activity on the
D
quality of water collected in roof water catchment systems at Stromboli Island (Italy). J. Geochem.
PT E
Explorat. 131, 28-36.
Mather, T.A., Pyle D.M., Oppenheimer, C., 2003. Tropospheric volcanic aerosol, in volcanism and the
CE
Earth’s atmosphere. In Volcanism and the Earth’s Atmosphere (eds. A. Robock and C. Oppenheimer). Geophysical Monograph 139 pp. 189–212.
AC
Moune, S., Gauthier, P.J., Gislason, S.R., Sigmarsson, O., 2006. Trace element degassing and enrichment in the eruptive plume of the 2000 eruption of Hekla volcano, Iceland, Geochim. Cosmochim. Acta, 70, 461–479, doi:10.1016/j.gca.2005.09.011.
Naughton, J.J., Lewis, V.A., Hammond, D., Nishimoto, D., 1974. The chemistry of sublimates collected directly from lava fountains at Kilauea volcano, Hawaii, Geochim. Cosmochim. Acta, 38, 1679–1690, doi:10.1016/0016-7037(74)90185-9.
Oppenheimer, C., 2003. Volcanic degassing. In Treatise on Geochemistry, vol. 3 (eds. R. L. Rudnick, H.
14
ACCEPTED MANUSCRIPT D. Holland and K. K. Turekian). Elsevier-Pergamon, Oxford, pp. 123–166.
Oskarsson, N., 1980. The interaction between volcanic gases and tephra: Fluorine adhering to tephra of the 1970 Hekla eruption, J. Volcanol. Geotherm. Res., 8, 251–266, doi:10.1016/03770273(80)90107-9.
Patrick, M., Harris, A.J.L., Ripepe, M., Dehn, J., Rothary, D.A., Calvari , S., 2007. Strombolian explosive
PT
styles and source conditions: insights from thermal (FLIR) video. Bull. Volcanol. 69, 769–784.
RI
Rizzo, A.L., Federico, C., Inguaggiato, S., Sollami, A., Tantillo, M., Vita, F., Bellomo, S., Longo, M., Grassa, F., Liuzzo, M., 2015. The 2014 effusive eruption at Stromboli volcano (Italy): Inferences from
SC
soil CO2 flux and 3He/4He ratio in thermal waters. Geophys. Res. Lett. 42, 22352243.
NU
Rose, W.I., 1977. Scavenging of volcanic aerosol by ash: Atmospheric and volcanologic implications,
MA
Geology 5, 621–624, doi:10.1130/0091-7613(1977)5<621:SOVABA>2.0.CO;2.
Scholl, M.A., Ingebritsen, S.E., 1995. Total and non-seasalt sulfate and chloride measured in bulk precipitation samples from the Kilauea volcano area, Hawaii. US Geol. Surv. Water-Resour. Invest.
PT E
D
Rep. 95–4001, 32 pp
Siegel, B.Z., Nachbar-Napai, M., Siegel, S.M., 1990. The contribution of sulfate to rainfall pH around
CE
Kilauea volcano, Hawaii. Wat. Air Soil Pollut. 52, 227–235.
Taddeucci, J., Scarlato, P., Capponi, A., Del Bello, E., Cimarelli, C., Palladino, D.M., Kueppers, U., 2012.
AC
High- speed imaging of Strombolian explosions: the ejection velocity of pyroclasts. Geophys. Res. Lett. 39, L02301.
Textor, C., Graf, H.-F., Timmreck, C., Robock, A., 2004. Emissions from volcanoes. In: Granier, C., Artaxo, P., Reeves, C. (Eds.), Emissions of Atmospheric Trace Compounds. Kluwer, Dordrecht, pp. 269–303.
Varekamp, J.C., Thomas, E., Germani, M., Buseck, P.R., 1986. Particle geochemistry of volcanic plumes
of
Etna
and
Mount
St.
Helens,
J.
Geophys.
Res.,
91,
12,233–12,248,
doi:10.1029/JB091iB12p12233.
15
ACCEPTED MANUSCRIPT
Von Glasow, R., Bobrowski, N., Kern, C., 2009. The effects of volcanic eruptions on atmospheric chemistry. Chem. Geol. 263, 131–142.
WHO, 2011. Guidelines for Drinking-water Quality. Fourth edition. WHO Press, Geneva. Switzerland, 564 pp.
PT
Zoller, W.H., Parrington, J.R., Phelan‐Kotra, J.M., 1983. Iridium enrichment in airborne particles from
AC
CE
PT E
D
MA
NU
SC
RI
Kilauea volcano: January 1983, Science, 222, 1118–1121, doi:10.1126/science.222.4628.1118.
16
ACCEPTED MANUSCRIPT Figure captions Fig. 1 Map of Stromboli island with location of sampling sites and main wind direction; Fig. 2 Rain gauge and its assemblage: a) picture of the island showing the plume moving downslope toward south-east and the FRT sampling site location; b) exploded view drawing of the rain gauge showing its different components; the perforated disks are used for supporting paper filters; c) axonometric view of the assembled rain gauge; d) front view of the rain gauge showing the container and the funnel; e) picture showing the rain sampler installed at COA station. In the centre, the orange
PT
PVC tube mounted on a metallic pole contains the rainwater sampling system illustrated in b and c; Fig. 3 Values of pH (a, b, c), TDS (d, e, f) and F concentrations (g, h, i) plotted vs distance from active
RI
vents, altitude and angle between the main propagation direction of the plume and the line
SC
connecting the vents with sampling sites. To avoid errors due to different sampling periods, plotted values are related to the temporal interval for which samples are available for all the sites (August 2015 – March 2016).
NU
Fig. 4 a) pH vs exposure time (days); b) TDS vs rain amount (mm);
Fig. 5 a) Langelier-Ludwig diagram. Samples having very similar compositions are sometimes not
MA
distinguishable in the graph because the related points could overlap; b) Temporal variation of Cl/SO4 ratio and F concentration in the COA site;
Fig. 6 Scatterplots correlating the molar abundances of cation and anion species in rainwater
D
samples: a) Cl vs SO4; b) SO4 vs Ca; c) Cl vs Na; d) Cl vs K. Black lines refer to plume’s chemical
PT E
composition, grey dotted lines represent the seawater ratio, and grey dotted-dashed lines represent the stechiometric ratio of the prevalent soluble mineralogical phases; Fig. 7 Trace elements composition range (coloured areas) of the different sampling sites (a) COA; b)
CE
FRT; d) DER). Black-dotted line represents the average composition of ash leachate (data from Bagnato et al., 2011), the blue line is the average composition of rainwater collected in roof catchment systems (Madonia et al., 2013a), and the blue-crossed diamonts are the safety limits for
AC
drinking waters established by the WHO (2011); Fig. 8 Enrichment factors (of the different sampling sites a) COA; b) FRT; c) DER) calculated for each element in rainwater, using Ti as reference element, with respect to ash composition (data from Allard et al., 2000; Bertagnini et al., 2003). Elements are ordered according to their mobility.
17
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT E
Fig. 1
18
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
Fig. 2
19
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
Fig. 3
20
AC
CE
PT E
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
Fig. 4
21
AC
CE
PT E
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
Fig. 5
22
PT E
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
Fig. 6
23
AC
CE
PT E
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
Fig. 7
24
AC
CE
PT E
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
25
AC
CE
PT E
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
Fig. 8
26
AC
CE
PT E
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
27
ACCEPTED MANUSCRIPT Table 1. Sample sites information Longitude WGS-84
Latitude WGS-84
COA DER FRT PDR SCA
15°.234189 15°.228126 15°.230482 15°.236543 15°.239173
38°.802776 38°.808216 38°.799138 38°.804936 38°.798198
Altitude (m a.s.l.) 73 25 198
Angle (°) 73 67 94
Distance from the crater (m) 2180 2070 1740
Sample type Gauge Gauge Gauge Single event Single event
AC
CE
PT E
D
MA
NU
SC
RI
PT
Site
28
ACCEPTED MANUSCRIPT
Site
Sampling date
Exposure time (day)
Rain (mm)
pH
EC
Na
K
Mg
Ca
Cl
8.03 4.51
100
12527 4.60
445 0.41
1501 0.68
447 2.13
22838 10.2
F
SO4
15 Jan 2015 15 Jan 2015
43
112
COA SCA
12 Mar 2015 12 Mar 2015
56 Single event
88
4.08 5.80
222 24.4
5.04 2.29
0.43 0.33
0.88 0.31
2.07 1.21
12.8 4.02
0.722 0.093
8.57 2.32
COA FRT
24 Apr 2015 24 Apr 2015
42 42
25 24
4.73 4.20
126 136
7.13 6.90
0.16 0.78
1.46 0.97
4.21 2.61
15.2 13.1
0.570 0.570
9.99 9.61
COA FRT PDR
21 May 2015 21 May 2015 21 May 2015
27 27 Single event
6.5 4.3
4.5 3.6 6.15
80.9
31.29 48.06 6.56
3.73 9.28 0.96
6.77 11.0 0.69
26.9 36.1 5.09
61.8 93.6 6.78
2.39 7.41 0.270
62.42 124 6.89
COA FRT
12 Aug 2015 12 Aug 2015
83 83
26 19
4.79 3.7
157
12.47 41.56
4.39 12.7
4.09 12.0
19.5 57.8
35.0 119
2.59 11.9
43.8 148
COA FRT DER
27 Jan 2016 27 Jan 2016 27 Jan 2016
168 168 168
468 333 386
3.71 3.38 4.14
70.4 120 72.4
4.49 6.15 7.34
0.00 0.97 0.72
0.77 1.47 1.12
1.71 3.89 2.52
10.2 15.3 14.3
0.505 1.11 0.378
6.91 15.9 7.88
COA FRT DER
18 Mar 2016 18 Mar 2016 18 Mar 2016
51 51 51
80 71 67
4.18 3.03 5.26
95.1 171.7 132.6
7.38 10.73 14.69
1.57 3.56 2.62
3.66 8.08 4.80
16.6 29.9 29.6
SC
RI
PT
SW COA
0.28 1.20 1.04
0.498
0.743 2.19 0.462
2981 6.16
11.56 26.98 14.06
AC
CE
PT E
D
MA
NU
Table 2. Concentration (in mg l-1) of major elements of rainwater samples collected on Stromboli Island between November 2014 and March 2016. EC is the Electrical Conductivity (S cm-1).
29
ACCEPTED MANUSCRIPT
Table 3. Concentration (g l-1) of minor and trace elements of rainwater samples collected on Stromboli Island between November 2014 and March 2016 Sampling date 15 Jan 2015
Zn 505
Ba 124
Al 210
Sb 113
Fe 33.9
Sr 18.5
Mn 11.8
B 3.12
Cu 10.0
Ni 2.45
Cr 0.400
As 0.769
Pb 0.810
Co 1.45
Rb 0.509
V 0.785
COA FRT
24 Apr 2015 24 Apr 2015
932 1317
0.363 324
222 272
66.1 140
43.2 52.9
27.9 32.9
18.7 23.8
16.4 15.5
3.31 10.3
3.80 6.65
0.853 1.58
0.817 1.13
0.299 1.01
1.16 0.881
0.233 0.856
0.206 0.583
COA FRT DER
27 Jan 2016 27 Jan 2016 27 Jan 2016
422 18.8 614
27.2 18.3 21.5
296 852 271
39.8 1.97 48.4
37.5 249 64.7
12.8 33.8 17.0
12.4 39.5 13.5
6.83 8.80 8.24
5.69 18.8 16.9
0.494 1.34 1.69
0.071 0.128 0.243
0.499 1.99 0.876
1.29 0.376 1.07
0.396 0.709 0.530
0.533 1.54 0.964
0.099 0.369 0.508
COA FRT DER
18 Mar 2016 18 Mar 2016 18 Mar 2016
143 40.7 1462
14.3 27.4 16.0
509 1410 96.5
24.5 1.69 65.9
272 996 28.7
27.5 50.8 27.9
36.3 78.5 33.5
14.5 27.7 22.0
8.19 40.3 6.52
1.81 6.54 2.23
0.136 0.334 0.150
0.612 2.42 1.01
0.521 2.02 0.183
0.810 1.45 1.25
0.575 2.02 1.24
0.299 1.75 0.286
AC
CE
PT E
D
MA
NU
SC
RI
PT
Site COA
30
ACCEPTED MANUSCRIPT Table 4: Linear correlation coefficients (r) between minor and trace element concentrations versus rainfall amount, duration of the sampling period and pH
D
pH -0.84 -0.64 -0.04 0.06 -0.77 -0.50 0.23 0.17 -0.76 -0.75 -0.70 -0.20 -0.54 0.46 -0.17 -0.71 -0.55 0.58 -0.34 -0.42 -0.82 -0.77 -0.60 -0.53 0.78
NU
SC
RI
PT
Sampling period (days) 0.05 0.02 -0.51 -0.34 -0.08 0.18 -0.80 -0.48 0.28 0.06 -0.15 0.06 -0.24 -0.49 -0.63 0.10 0.13 -0.46 -0.55 -0.41 0.19 0.21 -0.06 -0.31 -0.38
AC
CE
PT E
Al As B Ba Be Cd Co Cr Cs Cu Fe Li Mn Mo Ni Pb Rb Sb Se Sr Ti Tl U V Zn
Rain (mm) -0.02 -0.12 -0.59 -0.33 -0.16 0.00 -0.77 -0.53 0.17 0.00 -0.19 -0.03 -0.32 -0.39 -0.68 0.18 0.02 -0.39 -0.62 -0.54 0.10 0.15 -0.15 -0.31 -0.39
MA
Elements
31
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
Graphical abstract
32
ACCEPTED MANUSCRIPT Highlights Volcanic plume, sea spray aerosol and rainwater interact; Interaction occurs both in the atmosphere and after deposition;
AC
CE
PT E
D
MA
NU
SC
RI
PT
Chemistry of rain samples is conditioned by post-depositional processes.
33
Marianna Cangemi, Paolo Madonia, Rocco Favara PII: DOI: Reference:
S0377-0273(16)30448-6 doi: 10.1016/j.jvolgeores.2017.01.023 VOLGEO 6009
To appear in:
Journal of Volcanology and Geothermal Research
Received date: Revised date: Accepted date:
14 November 2016 27 January 2017 28 January 2017
Please cite this article as: Marianna Cangemi, Paolo Madonia, Rocco Favara , Chemical characterisation of rainwater at Stromboli Island (Italy): The effect of post-depositional processes. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Volgeo(2017), doi: 10.1016/j.jvolgeores.2017.01.023
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Chemical characterisation of rainwater at Stromboli Island (Italy): the effect of post-depositional processes Marianna Cangemia, Paolo Madoniaa*, Rocco Favaraa Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Palermo, via Ugo La Malfa 153, 90146 Palermo (Italy) *
Corresponding author, Email: [email protected]
PT
Abstract
Volcanoes emit fluids and solid particles into the atmosphere that modify the chemical composition
RI
of natural precipitation. We have investigated the geochemistry of Stromboli’s rainfall during the
SC
period from November 2014 to March 2016 using a network of a new type of sampler specifically designed for operations on volcanic islands. We found that most of the chemical modifications are due to processes occurring after the storage of rainwater in the sampling bottles. These processes
NU
include dissolution of volcanogenic soluble salts encrusting volcanic ash and a variable contribution of sea spray aerosol. Our data showed noticeably less scatter than has previously been achieved with
MA
a different sampling system that was more open to the atmosphere. This demonstrates the improved efficacy of the new sampler design. The data showed that post-depositional chemical alteration of rain samples dominates over processes occurring during droplet formation ad precipitation. This has
PT E
D
important implications for the calculation of fluxes of chemicals from rainfall in volcanic regions.
Keywords
1.
CE
Rainwater chemistry; sea spray; plume; volcanic ash; Stromboli; post-depositional processes.
Introduction
Volcanoes are important sources of gases, solid particles and aqueous acid droplets emitted at high
AC
temperature into the atmosphere (Cadle, 1980; Oppenheimer, 2003; Mather et al., 2003; Gerlach, 2004; Textor et al., 2004; Von Glasow et al., 2009). Atmospheric precipitations represent a significant mechanism for maintaining the balance between sources and sinks of particulate, providing a natural scavenging of the chemical dispersed into the atmosphere also in terms of acidification of water (Gonzáles and Aristizábal, 2012). The interaction between volcanic activity and rain chemistry has been investigated worldwide, and among the others at Kilauea, Hawai (Siegel et al. 1990; Scholl and Ingebritsen 1995), Soufriere Hills volcano, Monserrat (Edmonds et al. 2003), and in Italy at Mt Etna (Bellomo et al. 2003), Vulcano island (Liotta et al. 2006; Madonia and Liotta, 2010), Stromboli island (Bellomo et al., 2003; Liotta et al., 2006; Madonia et al., 2013), and Mt Vesuvius (Madonia and Liotta,
1
ACCEPTED MANUSCRIPT 2010). Rain droplets interact with volcanic products giving rise to rainwater with a strong chemical fingerprint of the dissolution of soluble ions and gases (Bellomo et al., 2003; Liotta et al. 2006; Madonia and Liotta, 2010). In addition, volcanic aerosol can act as cloud condensation nuclei promoting cloud formation. The deposition of precipitations chemically altered by the interaction with volcanogenic volatiles and particles impacts the surrounding environment, damaging vegetation (Delmelle et al. 2002), changing the soil chemistry (Delmelle et al. 2003) or affecting water supply systems (Cuoco et al., 2013a, b;
PT
Madonia et al., 2013).
During the last decades several studies focused on the chemistry of rainfalls in volcanic areas have
RI
been carried out on Italian active and quiescent volcanoes, because these are inserted into heavy
SC
urbanized contexts and the potential impacts on human health are highly significant (Bellomo et al., 2003; Liotta et al., 2006; Madonia and Liotta 2010; Madonia et al., 2013). Among these areas, Stromboli is particularly relevant because of both its continuous volcanic activity and the presence of
NU
a very populous human community, especially during the touristic season. A previous work on the geochemistry of Stromboli’s rainwater was conducted by Liotta et al. (2006).
MA
The authors investigated the chemical and isotopic compositions of the precipitation between October 2003 and October 2005, finding no direct contribution of volcanogenic H or O. The chemical composition of rainwater close to the coastline was found influenced by the sea aerosol, whereas the
D
chemical effects due to the volcanic activity were found in those samples collected closer the summit
PT E
vents. Moreover, the S/Cl, Cl/F, and S/F molar ratios were consistent with those measured in the volcanic plume. Minor and trace element concentrations in rainwater were not considered during this study and, more in general, in other researches in Mediterranean volcanic islands. The reason is
CE
that the protocols for collecting rainwater samples for minor and trace element determination are based on Büchner type funnels with a single sampling session covering a time span of about 15 days (Calabrese et al., 2011). This procedure does not fit the economical and logistic requirements for
AC
studies in insular areas, like Stromboli, where it is really difficult ensuring a so frequent and constant field activity.
Following these premises we designed a new type of rain sampler for the determination of minor and trace elements, able to properly collect samples over several months, and tested it at Stromboli island. We choose this location because previous studies characterised the chemical composition of volcanic ash (Bagnato et al., 2011) and rainwater collected both directly from the atmosphere (major elements only; Liotta et al., 2006) and harvested by roof catchment systems (Madonia et al., 2013). Moreover, during our sampling campaign, from November 2014 to March 2016, Stromboli was experiencing a period of different volcanic activity with respect to the previous study by Liotta et al.
2
ACCEPTED MANUSCRIPT (2006), giving us the opportunity to look for its possible proxies in the geochemical characterisation of precipitation.
2. Volcanological setting Stromboli, the north-easternmost island of the Quaternary Aeolian volcanic arc off the northern coast of Sicily (Southern Italy), is a stratovolcano that rises from a depth of ~2000 m b.s.l. in the Tyrrhenian sea with a subaerial portion 924 m high (Fig. 1).
PT
It is characterised by the so called “Strombolian activity”, consisting of passive magma degassing alternated to short (a few to few tens of s) 100- to 200-m high scoria-rich jets produced by explosions
RI
of variable energy every 10–20 min (Patrick et al., 2007; Andronico et al., 2008; Taddeucci et al.,
SC
2012). This activity take place from vents lying on the NE-SW-trending crater terrace, located at about 750 m above sea level (a.s.l.) in the upper portion of the Sciara del Fuoco. The volcanic plume continuously emitted by the vents is mainly composed of water vapour (over 50
NU
wt%), with CO2 and SO2 as other principal gaseous species; HCl and HF are also emitted as secondary
3.
MA
components (Allard et al., 1994).
Sampling and analytical methods 3.1
Sampling methods
D
As previously introduced we realised a new sampler for collecting rainwater samples for trace
PT E
element analysis, designed for properly working over several months: the main issues in using the normal Büchner type funnels (Calabrese et al., 2011) are the reduced volume of the sampling bottle and its insufficient isolation from the atmosphere. For solving these problems, we designed and
CE
realized a rain sampler where the funnel diameter/sampling bottle volume ratio allows collecting up to 520 mm of rain; based on literature data (Cicala, 2000), this height corresponds to the total precipitations falling at Stromboli during several months of the autumn-winter wet season.
AC
Considered the extended duration of the sampling period, we adopted an array of 3 sequential filters (Fig. 2) instead of the single filter usually equipping the Büchner type funnels: in this way we reduced possible evaporation losses and contamination due to the input of fine particulate into the sampling bottle. About the construction details, our system consists of a 2 l Nalgene bottle over which a composite TeflonTM funnel, with a diameter of 75 mm, is fasten. The funnel can be separated in each component for easy and total cleaning. A sequence of three pierced disks with a downward decreasing hole diameter (5, 4, 3 mm respectively) is inserted inside the funnel. Sheets of filter paper (Whatman® Cellulose quantitative filter paper, ashless, grade 40) are put between each pierced disks
3
ACCEPTED MANUSCRIPT with the aim of avoiding the clog of the holes, preventing ash entering the sampling bottle and consequently minimizing the prolonged interaction between volcanic ash and rainwater after collection. Such a system is also useful for reducing evaporation of water after the collection: in fact, in normal Büchner funnel there is only one filter that is strongly heated by the incident solar radiation. In our system the multiple disks work as a radiative shield, limiting heat transfer to the interior of the bottle and evaporation of the collected water. Moreover, the rain sampling system is put into a plastic case, with an air gap between the case and the bottle for further reducing radiative
PT
heating. The whole system is mounted on a metallic pole at about 1.5 m above the ground, ensuring the horizontality of the catchment area of the funnel. All the equipments were prewashed with a 2%
RI
nitric acid (Merck Suprapur) solution, and then rinsed several times with de-ionized (MilliQ) water,
SC
dried under hood, packed in double clean plastic bags, and zipped until exposure in the field. At the end of the sampling period, the sampling bottle was substituted with a new clean one, capped and transported in the laboratory for analysis.
NU
A total of 3 rain sampling systems were installed at different distance from the crater and with different directions respect to the dominant winds (Fig. 1; Table 1). The first prototype was installed
MA
at the COA volcanological observatory in November 2014 and, after a preliminary test on its performances, we built up the sampling network with another system installed at FRT (in the same location of Liotta et al., 2006) in January 2015. We completed the network in August 2015 with a
D
third system in the Piscità area (DER, Fig. 1). The sampling sites are distributed both at variable
PT E
distances from the summit vents and at variable angular distances with respect to the main volcanic plume propagation direction, allowing us to estimate the possible interactions between volcanic activity and rainfalls. Moreover, 2 single rain events were collected at SCA (close to Stromboli’s
respectively.
CE
harbour) and PDR (few hundreds of metres northernmost of COA) in March and May 2015,
3.2 Analytical methods
AC
Chemical-physical parameters (pH and electric conductivity) were measured directly in the field using a Thermo Orion instrument and a 8102BN Ross combination pH electrode. Water samples used for the determination of dissolved major and trace elements were first filtered using 0.45 µm Millipore MF filter and then: a) collected in LD-PE (low-density polyethylene) bottles for major element analyses, acidifying to ca pH 2 the aliquot destined to cation determination with HCl; b) collected in PP (polypropylene) bottles for trace element analyses, acidified to pH ca 2 with ultrapure concentrated HNO3. Untreated aliquots were stored for alkalinity determinations, made via titration with HCl (0.1 N). Electric conductivity and pH were measured in the field using Orion instruments equipped with Hamilton electrode (pH). Major ions were determined by ionic chromatography using
4
ACCEPTED MANUSCRIPT Dionex columns AS14 and CS12 for anions and cations respectively. Major element data are presented in Table 2. The trace elements, were analysed by ICP-MS (Agilent 7500ce) equipped with a Micromist nebulizer, a Scott double pass spray chamber, a three-channel peristaltic pump, an auto sampler (ASX-500, Cetac), and an octopole reaction system (ORS) to remove the interferences. The mass spectrometer was calibrated with a multi-element standard solution with 11 calibration points. A multi-element standard solution was prepared by dilution from single and multi-element stock solution,
PT
respectively at 1.000 mg/L and 100 mg/L. Multi-element stock calibration solutions were always prepared freshly in 10mL polyethylene tubes by dilution using 1% nitric acid for trace analysis as
RI
diluent.
SC
The sensitivity variations were monitored using 103Rh, 115In, and 185Re, internal standards were added directly online by an appropriate device that mixes the internal standard solution into the sample just before the nebulizer. The concentration of the internal standard was about 10 μg/L in all sample
NU
and calibration solutions. The precision of the analysis was checked by running five replicates of every standard, and the sample was always within ±10%. Data accuracy was evaluated by analysing
MA
standard reference materials (Spectrapure Standards SW1 and 2, SLRS4, NIST 1643e, Environment Canada TM 24.3, and TM 61.2) for each analytical session, and the error for each element was <15%.
Results
Chemico-physical parameters and concentration of major elements
PT E
4.1
D
4.
In order to evaluate the spatial variations of pH and TDS we plotted these values against three different parameters expressing the possible effect of volcanic activity, and in particular of the
CE
interaction with the volcanic plume. These are the planar, vertical and angular distances of the sampling sites with respect to the summit vents, expressed in terms of relative distances, altitude of the sites, and angles formed between the propagation direction of the plume and the line connecting
AC
the vents with the sampling point. Although only 3 points are available, thus not giving the opportunity of calculating statistically significant relationships, useful qualitative information can be retrieved. These relationships are illustrated in Fig. 3a-I, showing the variations of pH (a-c) and TDS (d-f) with distance from the summit vents, altitude and angle with respect to plume direction. The pH of rainwater was highly variable, ranging from 3.0 to 5.3 (Table 2); as shown in Fig. 3a-c, correlations are found with altitude and angle, while the relationship with distance is not so clear. The closest site to the summit vents (FRT), at highest altitude and the lowest angular distance from the propagation direction of the plume, was characterised by the most acid water (pH range 3.0-4.2). On the contrary, the highest pH values were recorded at DER (range 4.1-5.3), that is the lowest altitude and at the
5
ACCEPTED MANUSCRIPT maximum angular distance from the propagation direction of the plume. The COA site shows intermediate values. Similar relationships are found for total dissolved solids (TDS), with highest values at FRT and lowest at DER with COA in an intermediate position (Fig. 3d-f). About the temporal variations of these parameters, it is difficult discriminating between their possible dependence on duration of sampling intervals or total amounts of precipitation, because these two are also strongly correlated (r2= 0.97). Anyway, Fig. 4a highlights a progressive decrease of
PT
pH following the increase of the sampling period (r= -0.65); two single rain events (PDR and SCA) are also reported for comparison. Similarly, TDS negatively correlates with rainfall amount (Fig. 4b).
RI
Major element concentrations are listed in Table 2; electroneutrality balance is not assured because
SC
not all the ionic species were determined but only those of interest for the aim of the present study. In the modified Langelier-Ludwig diagram, shown in Fig. 5a, the sampled rainwaters are aligned along a compositional trend produced by different mineralization processes. The two end members are sea
NU
spray aerosol and a sulphate-dominated term, typical of the Stromboli volcanic environment (Liotta et al., 2006). Groundwater sampled into wells located in the coastal area of the island (Grassa et al.,
MA
2008) lies along this curve, close to the seawater composition. A difference is observed among our and previous (Liotta et al., 2006) data. The former show a good alignment, whereas the latter are much more scattered, especially for compositions close to the volcanic end-member. Some intra-
D
annual compositional fluctuations are also observed and shown in Fig. 5b, related to the COA site for
PT E
which the longest time series, covering more than a complete year, is available. We plotted Cl/SO4 ratio and F concentrations vs time: the two variables are anti-correlated and indicate an increase in fluoride concentration during the dry season, coupled to a relative enrichment of SO4 with respect to
CE
Cl (see Table 2). Fluoride also varies in space, as shown in Fig. 3g-I, as it follows the same pattern of TDS: the smallest are the planar (with the exception of COA) and vertical distances from the summit vents and the angle formed between the propagation direction of the plume and the line connecting
AC
the vents with the sampling points, the highest is F concentration. It is relevant that F concentration often exceeds the safety limits for drinking water, fixed at 1.5 mg l-1 by the World Health Organization (WHO, 2011), at FRT and COA, with a maximum value of 11.9 mg l-1 measured in the August 2015 sample at FRT. The scatterplots of Fig. 6 (a-d) compare pairs of concentrations of the most representative ions in rainwater (Na, K, Ca, Cl, SO4) with respect to those of volcanic plume (Allard et al., 2000), Mediterranean seawater and the mineralogical phases by whose dissolution these can derive. As a general remark, all the rainwater samples are comprised in an area delimited by the lines representative of the possible chemical end-members.
6
ACCEPTED MANUSCRIPT 4.2
Minor and trace metal concentrations
Metal concentrations are listed in Table 3 and illustrated in Fig. 7 in roughly order of relative abundances. These range from about 0.005 g l-1 for ultra-trace elements (e.g. Be, Cs, Mo, U) up to ≥1000 g l-1 for minor elements (Al, Li, Zn). The concentration patterns are similar to those found by Madonia et al. (2013) in rainwater collected in roof catchment systems and Bagnato et al. (2011) in ash leachate, with intermediate absolute values between these end members. The safety limits fixed by the WHO for drinking water are also reported in Fig. 7. As shown in the
PT
diagrams, all the concentrations of the determined elements fall below these limits, with the exception of Sb that exceeds its threshold in all the sites.
RI
As expected for metals in aqueous solution, the concentration of quite all the determined elements
SC
negatively correlate with pH (Table 4). Exceptions are represented by Zn and Mo. For better constraining the relative amount of each element in rainwater with respect to its concentration in volcanic ash, we calculated the enrichment factor (EF) for each X element, following
NU
the relation (Zoller et al., 1983):
EFash = (X/Ti)rain/(X/Ti)ash; EFash leachate = (X/Ti)rain/(X/Ti)ash leachate
MA
We used Ti as reference element due its low volatility under magmatic conditions and low mobility during weathering processes (Aiuppa et al., 2000). EFs are shown in Figure 8 ordered according to their mobility. Sampling sites show enrichments in Mo, Zn, Tl, As, Cu and Ni, with respect to ash
D
(Allard et al., 2000; Bertagnini et al., 2003). Elements like Ba, Sr, Rb, Al, Fe, Li, and Cr have EFs values
5.
Discussion
PT E
below the unity.
CE
Due to the logistic and economic reasons discussed in the introduction, data about minor and trace elements in rainwater from volcanic islands are rare worldwide and absent from the Mediterranean.
AC
Ours are the first available data from a Mediterranean volcanic island, thanks to a sampling system specifically designed for collecting multi-months samples. A first controversial question, already arisen from previous studies devoted to the geochemical characterization of rainwater in volcanic areas, is whether the chemical composition of the sampled precipitation is driven by interaction processes occurring in the atmosphere, or the sample is chemically modified after the collection in the sampling bottle. The question is not captious, because if the second hypothesis is true, no quantitative inferences about the flux of volcanogenic elements in the atmosphere can be extracted using as a proxy the chemical composition of rainwater. In a study based at Vulcano Island, Madonia and Liotta (2010) demonstrated that the interaction with acidic gases, emitted by a fumarolic field, is able to strongly modifying the chemical fingerprint of
7
ACCEPTED MANUSCRIPT rain samples after collection. These authors exposed to the atmosphere a bottle containing MilliQ water in a location few hundreds of meters downwind of the fumarolic field of La Fossa cone, finding over 20 milliequivalents of dissolved sulphate after 2 months of exposition. A similar behaviour is found in the chemistry of the samples collected during our study. As shown in Fig. 4a, the pH of the collected rainwater negatively correlates with the duration of the sampling period, suggesting that the water collected in the bottle progressively reacts with volcanic acidic gases (Delmelle, 2003) continuously emitted from the summit vents. In fact, as illustrated in Fig. 3a-c,
PT
the closest a sampler is located to areas affected by interactions with volcanic plume (FRT) the highest is the acidity of the collected rainwater. This must not be intended as a planar distance,
RI
because the graphs indicate that the direction of the plume (Fig. 3c) and the elevation proximity (Fig.
SC
3b) are more effective in controlling pH. In the same way, an opposite correlation is found for TDS (Fig. 3d-f) because the most acid is the water the higher is its capability of scavenging chemicals from ash particles. The fingerprint of this process is found in the exponential decrease of TDS with the
NU
increase of rain amount (Fig. 4b). Chemical composition of rainwater is controlled by the leaching of soluble salts, both coated on the surface of volcanic particles accumulated into the sampler funnel
MA
(outside the multiple filter system) or coming from the deposition of sea spray. When the volume of water accumulated into the bottle is low, the dissolution of these particles leads to a higher content of dissolved ions. Conversely, when the water volume is higher, the final ion concentration will be
D
lower. These observations are confirmed by the minor and trace element behaviour. As shown in
PT E
Table 4, the concentrations of the various species negatively correlate with both amount of rainfall and duration of sampling periods.
The novel inference from our study is that the chemistry of rainwater sampled in active volcanic
CE
areas with wet and dry rain gauges is primarily influenced by reactions occurring after the collection of rainwater into the samplers. We were able to obtain this data because we collected for the first time rainwater samples representative of the total amount of precipitation in several months. Lot of
AC
caution should then be adopted for extracting quantitative inferences on the atmospheric emission of volcanogenic elements based on their concentration in rainwater samples. Although quantitative flux models seem difficult to be developed, interesting qualitative information concerning the origin of the elements and their variation with time can be extracted from the chemical composition of rainwater collected at Stromboli. A double origin for the dissolved ions, sea spray and volcanogenic compounds, is supported by the Langelier-Ludwig diagram (Fig. 5a). It is noteworthy that SO4 is enriched with respect to Cl during the spring-summer season, as well as fluorine (Fig. 5b), indicating a stronger volcanic fingerprint during these months. This effect could be driven by the coaction of two different processes: i) in winter
8
ACCEPTED MANUSCRIPT winds are stronger and sea waves higher, producing more sea spray that is transported and deposited into the funnels with respect to the spring-summer season; ii) during winter atmospheric relative humidity is higher than in warm season. Since acidic species are strongly hydrophilic these are quickly removed from the plume, due to the interaction with the atmospheric water droplets. As a consequence, the residual volcanic plume interacting with rainwater accumulated into the samplers will be depleted in acidic species. This last consideration opens a question about a possible implication for remote sensing of volcanic plume composition, since this scavenging may also be
PT
active during these measurements.
The dissolution of the volcanogenic coating material encrusted on volcanic ash is the other major
RI
source for chemicals found in rainwater. This material is mainly composed of CaSO4 (anhydrite or
SC
gypsum), halite and sylvite, formed during the reactions between acidic gases (mainly SO2g and HClg), abundant in the plume, and the volcanic particles (Rose, 1977; Oskarsson, 1980; Hinkley and Smith, 1982; Delmelle et al., 2007; Bagnato et al., 2011; Barone et al., 2016). These minerals have also been
NU
found in volcanic ash from Mt Etna (Barone et al., 2016). These dissolution processes are clearly identified in the binary diagrams of Fig. 6. In the SO4 vs Cl plot (Fig. 6a) rainwater is comprised in a
MA
band delimited by the lines representative of plume and seawater compositions. The Ca vs SO4 (Fig. 6b), Na vs Cl (Fig. 6c), and K vs Cl (Fig. 6d) diagrams evidence that the dissolution of anhydrite, halite and sylvite and the interactions with sea spray and plume explain the ratios found among these
D
compounds in the analysed rain samples.
PT E
The comparison between our data and those collected by Liotta et al. (2006) gives insights about the efficiency of our sampling system. Since that study was aimed to retrieve chemical and isotopic information from the collected samples, different samplers were used. These were made of 25-l
CE
cylinder shaped PET tanks with a 30-cm-diameter PET funnel, filled with Vaseline oil (250 cc) in order to avoid evaporation of the collected rain (more details about this system can be found in the paper by Madonia and Liotta, 2010). It is noteworthy that such a system is not equipped with any filter able
AC
to avoid the input of solid particulate inside the collection tank. A different behaviour is shown by our data in the Langelier-Ludwig diagram (Fig. 5) with respect to those from the previous study by Liotta et al. (2006), with the former better aligned following the trend having seawater and volcanic end members with respect to the latter. This difference suggests that our sampling system works better, because the multi-paper filters mounted on the funnel avoid the entrapment of solid volcanic particulate into the sampling bottles. The leaching of this particulate by the rain collected into the samplers could be responsible of the major variability of the chemical composition found by Liotta et al. (2006). Truthfully, a different explanation could be attributed to a more variable volcanogenic source during the years 2003-2005 (data from Liotta et al., 2006) with respect to a most recent
9
ACCEPTED MANUSCRIPT period (2014-2016, our data). Both sampling periods followed eruptive phases, occurred at Stromboli between December 2002 and July 2003 and August and November 2014, respectively. The two eruptions presented different styles: the 2002-03 event was energetic and showed an evident overpressurization of the shallow fracture system that led to a paroxysmal explosion on 5 April 2003, while the most recent event was more “quiet” due to a lesser amount of volatiles involved (Rizzo et al., 2015, and references therein). This difference in the degassing style seems being reflected in the two datasets. A more turbulent regime characterised the period 2003-05, leading to variable ratios
PT
between volcanogenic and sea spray elements. By contrast, a more invariant volcanogenic source dominated the period 2014-16, resulting in a quite constant chemical composition of rainwater.
RI
A special attention deserves F concentration. This element often exceeds the safety limit established
SC
by the WHO for drinking water in COA and FRT sites (Table 2), suggesting that the collection and storage of rainfalls for sanitary and alimentary purposes could imply human health issues. Nevertheless, the previous study by Madonia et al. (2013) on the quality of water collected in roof
NU
water catchment systems (RWCS) at Stromboli indicated that the RWCS water is generally close to the optimal concentration of 1 mg/l. This discrepancy can be attributed to the post-depositional
MA
reactions between the volcanic plume and the rainwater collected in the samplers; the higher F concentrations found in the present study are reasonably due to the small volume of water contained in the sampling bottles, strongly affected by the interaction with the plume (the highest F
D
concentrations are found in FRT site, which is the closest to the summit vents). Conversely, the water
PT E
stored in closed cisterns far away from the crater terrace is less contaminated by volcanogenic fluorine.
About trace elements, a first evidence is a lower content of Sr and Rb in rainwater with respect to
CE
samples collected in roof catchment systems (Madonia et al., 2013): this is likely due to the interaction between the harvested rainwater and the surface of the cisterns where these are stored, treated with lime. More generally, since the trace element pattern is similar to that found in ash
AC
leachate (Bagnato et al., 2011), leached volcanic ash is the most probable source for the elements found in rainwater.
Calculations of enrichment factors (Fig. 8) show that rainwater are enriched in those elements typical of ash that had interacted with the plume, like Mo, Zn, Tl, Cu, As, and Ni. These elements, present on ash surface under the form of halides and S compounds (e.g. CuSg, ZnCl2g, Krauskopf, 1964; Naughton et al., 1974; Moune et al., 2006; Varekamp et al., 1986), are easily leached by the rain. On the contrary, EFs below the unity characterize the so called refractory elements, less mobile during chemical weathering processes.
10
ACCEPTED MANUSCRIPT A final consideration concerns the possible issue for human health related to high concentrations of potentially toxic elements present in rainwater. As shown in Fig. 7 the only element exceeding the safety limit established by the WHO is Sb in all the sampling sites. However, as already discussed for F, this apparent high concentration could be caused by the post-depositional interaction with the volcanic plume in the rain samplers, which is not effective for water collected in RWCS; the comparison with the data by Madonia et al. (2013) confirms that rainwater harvested in cisterns is
6.
PT
always below the safety limits.
Conclusions
RI
First of all, our findings are relevant from a methodological point of view. Since the most of the
SC
water-rock interaction processes occur between rainwater and volcanic particles accumulated into the funnel, the use of wet and dry sampler should be mandatory in studies devoted to the chemical characterisation of natural precipitation in volcanic areas. Sampling systems based on the collection
NU
of bulk samples seem to be not adequate for such kind of researches, because it is impossible to discriminate between post-depositional scavenging of ash (leaching of solid particles accumulated
MA
into the funnels) and water-rock interaction processes occurring during rain droplets formation and precipitation (dissolution of gaseous and solid particles from volcanic plumes into the atmosphere). Although the two processes can not be distinguished, chemistry of rain collected in bulk samplers
D
could give useful, even though qualitative, information on the composition of a volcanic plume. In
PT E
this sense a bulk rain sampler could work as a passive trap for volcanic acidic species contained in the plume, which are tracers of the volcanic activity state. The more a volcano is active the more its plume is enriched in acidic species. At the same time, the more a plume is acidic the more it will be
CE
able to modify the chemical composition of the collected rain.
Acknowledgments
AC
This work was partially funded and realised in the behalf of the FSE postgraduate specialisation course “Valutazione e monitoraggio dei rischi ambientali naturali” (Regione Siciliana, Asse IV, Capitale Umano). We wish to thank our colleagues of INGV Paolo Cosenza and Giuseppe Riccobono, who improved the original project of the sampling system with their suggestions and materially realized the devices. We are also grateful to two anonymous reviewers and the editor Prof. Heidy M. Mader who gave us useful suggestions for improving this paper.
References Aiuppa, A., Allard, P., D’Alessandro, W., Michel, A., Parello, F., Treuil, M., Valenza, M., 2000. Mobility
11
ACCEPTED MANUSCRIPT and fluxes of major, minor and trace metals during basalt weathering at Mt. Etna volcano (Sicily). Geochim. Cosmochim. Acta, 64, 1827–1841, doi:10.1016/S0016-7037(00)00345-8.
Allard, P., Aiuppa, A., Loyer, H., Carrot, F., Gaudry, A., Pinte, G., Michel, A., Dongarrà G., 2000. Acid gas and metal emission rates during longlived basalt degassing at Stromboli volcano. Geophys. Res. Lett. 27(8), 1207–1210, doi:10.1029/1999GL008413.
RI
degassing budget of Stromboli volcano, Nature, 368, 326330.
PT
Allard, P., Carbonelle, J., Metrich, M., Loyer, H., Zettwoog, P., 1994. Sulphur output and magma
SC
Andronico, D., Corsaro, R.A.. Cristaldi, A., Polacci, M., 2008. Characterizing high energy explosive eruptions at Stromboli volcano using multidisciplinary data: An example from the 9 January 2005
NU
explosion. J. Volcanol. Geotherm. Res. 176, 541–550.
Bagnato, E., Aiuppa, A., Andronico, D., Cristaldi, A., Liotta, M., Brusca, L., Miraglia, L., 2011. Leachate
MA
analyses of volcanic ashes from Stromboli volcano: A proxy for the volcanic gas plum composition? J. Geophys. Res. 116:D17204, doi:10.1029/2010JD015512,2011
D
Barone, G., Mazzoleni, P., Corsaro, R.A., Costagliola, P., Di Benedetto, F., Ciliberto, E., Gimeno, D.,
PT E
Bongiorno, C., Spinella, C., 2016. Nanoscale surface modification of Mt. Etna volcanic ashes. Geoch. Cosmochim.
CE
Bellomo, S., D'Alessandro, W., Longo, M., 2003. Volcanogenic fluorine in rainwater around active
AC
degassing volcanoes: Mt. Etna and Stromboli Island, Italy. Sci. Total Environ. 301, 175–185.
Bertagnini, A., Métrich, N., Landi, P., Rosi M., 2003. Stromboli volcano (Aeolian Archipelago, Italy): An open window on the deep feeding system of a steady state volcano, J. Geophys. Res. 108(B7), 2336, doi:10.1029/2002JB002146.
Cadle R.D., 1980. A comparison of volcanic with other fluxes of atmospheric trace gas constituents. Rev. Geophys. Space Phys. 18, 746–752.
Calabrese, S., Aiuppa, A., Allard, P., Bagnato, E., Brusca, L., D’Alessandro, W., Parello, F., 2011. Atmospheric sources and sinks of volcanogenic elements in a basaltic volcano (Etna, Italy). Geoch.
12
ACCEPTED MANUSCRIPT Cosmochim. Acta 75/23, 7401-7425, doi: 10.1016/j.gca.2011.09.040.
Cicala, A., 2000. Guida alla Meteorologia delle Isole Eolie. Aldo Natoli Editore, Lipari, 149pp.
Cuoco, E., Tedesco, D., Podera, R.J., Williams, J.C., De Francesco, S., Balagizi, C., Darrah, T.H., 2013a. Impact of volcanic plume emissions on rain water chemistry during the January 2010 Nyamuragira eruptive event: Implications for essential potable water resources. J. Hazard Mater.
PT
581
RI
Cuoco, E., Spagnuolo, A., Balagizi, C., De Francesco, S., Tassi, F., Vaselli, O., Tedesco, D., 2013b.
SC
Impact of volcanic emissions on rainwater chemistry: The case of Mt Nyiragongo in the Virunga volcanic region (DRC). J. Geochem.
79.
NU
Delmelle, P., Stix, J., Baxter, P.J., Garcia-Alvarez, J., Barquero, J., 2002. Atmospheric dispersion, environmental effects and potential health hazard associated with the low-altitude gas plume of
MA
Masaya volcano, Nicaragua. Bull. Volcanol. 64, 423–434.
Delmelle, P., 2003. Environmental impacts of tropospheric volcanic gas plumes. In: Oppenheimer, C.,
PT E
D
Pyle, D.M., Barclay, J. (eds). Volcanic degassing, vol. 213. Geol. Soc. Spec. Publ. Lond. pp 381–399.
Delmelle, P., Lambert, M., Dufrene, Y., Gerin, P., Oskarsson, N., 2007. Gas/aerosol‐ash interaction in volcanic plumes: New insights from surface analyses of fine ash particles, Earth Planet. Sci. Lett. 259,
CE
159–170, doi:10.1016/j.epsl.2007.04.052.
Edmonds, M., Oppenheimer, C., Pyle, D.M., Herd, R.A., 2003. Rainwater and ash leachate analysis as
AC
proxies for plume chemistry at Soufriere Hills Volcano, Montserrat. In: Oppenheimer C, Pyle DM, Barclay J (eds). Volcanic degassing. Geological Society, London, pp 203–218
Gerlach, T.M., 2004. Volcanic sources of tropospheric ozone-depleting trace gases. Geochem, Geophys. Geosyst. 5 (Q09), 007.
Gonzáles, C.M., Aristizábal, B.H., 2012. Acid rain and particulate matter dynamics in a mid-sized Andean city: The effect of rain intensity on ion scavenging. Atmospheric Enviorn. 60, 164-171.
13
ACCEPTED MANUSCRIPT Grassa, F., Inguaggiato, S., Liotta, M., 2008. Fluid geochemistry of Stromboli. In: Calvari S, Inguaggiato S, Puglisi G, Ripepe M, Rosi M (eds) The Stromboli volcano: an integrated study of the 2002–2003 eruption. Am. Geophys. Union, Geophys. Mon.
Hinkley, T.K., Smith, K.S., 1982. Leachate chemistry of the tephra from the May 18 1980 eruption of Mount St. Helens, Eos Trans. AGU, 63(45), 1143.
PT
Krauskopf, K.B., 1964. The possible role of metal compounds in ore genesis, Econ. Geol., 59, 22–45,
RI
doi:10.2113/gsecongeo.59.1.22.
SC
Liotta, M., Brusca, L., Grassa, F., Inguaggiato, S., Longo, M., Madonia, P., 2006. Geochemistry of rainfall at Stromboli volcano (Aeolian Islands): isotopic composition and plume-rain interaction.
NU
Geochem. Geophys. Geosyst. 7Q07006. doi:10.1029/2006GC001288
Madonia, P., Liotta, M., 2010. Chemical composition of precipitation at Mt Vesuvius and Vulcano
MA
Island, Italy: volcanological and environmental implications. Environ. Earth Sci. 61, 159-171.
Madonia, P., Cangemi, M., Bellomo, S., D’Alessandro, W., 2013. Influence of volcanic activity on the
D
quality of water collected in roof water catchment systems at Stromboli Island (Italy). J. Geochem.
PT E
Explorat. 131, 28-36.
Mather, T.A., Pyle D.M., Oppenheimer, C., 2003. Tropospheric volcanic aerosol, in volcanism and the
CE
Earth’s atmosphere. In Volcanism and the Earth’s Atmosphere (eds. A. Robock and C. Oppenheimer). Geophysical Monograph 139 pp. 189–212.
AC
Moune, S., Gauthier, P.J., Gislason, S.R., Sigmarsson, O., 2006. Trace element degassing and enrichment in the eruptive plume of the 2000 eruption of Hekla volcano, Iceland, Geochim. Cosmochim. Acta, 70, 461–479, doi:10.1016/j.gca.2005.09.011.
Naughton, J.J., Lewis, V.A., Hammond, D., Nishimoto, D., 1974. The chemistry of sublimates collected directly from lava fountains at Kilauea volcano, Hawaii, Geochim. Cosmochim. Acta, 38, 1679–1690, doi:10.1016/0016-7037(74)90185-9.
Oppenheimer, C., 2003. Volcanic degassing. In Treatise on Geochemistry, vol. 3 (eds. R. L. Rudnick, H.
14
ACCEPTED MANUSCRIPT D. Holland and K. K. Turekian). Elsevier-Pergamon, Oxford, pp. 123–166.
Oskarsson, N., 1980. The interaction between volcanic gases and tephra: Fluorine adhering to tephra of the 1970 Hekla eruption, J. Volcanol. Geotherm. Res., 8, 251–266, doi:10.1016/03770273(80)90107-9.
Patrick, M., Harris, A.J.L., Ripepe, M., Dehn, J., Rothary, D.A., Calvari , S., 2007. Strombolian explosive
PT
styles and source conditions: insights from thermal (FLIR) video. Bull. Volcanol. 69, 769–784.
RI
Rizzo, A.L., Federico, C., Inguaggiato, S., Sollami, A., Tantillo, M., Vita, F., Bellomo, S., Longo, M., Grassa, F., Liuzzo, M., 2015. The 2014 effusive eruption at Stromboli volcano (Italy): Inferences from
SC
soil CO2 flux and 3He/4He ratio in thermal waters. Geophys. Res. Lett. 42, 22352243.
NU
Rose, W.I., 1977. Scavenging of volcanic aerosol by ash: Atmospheric and volcanologic implications,
MA
Geology 5, 621–624, doi:10.1130/0091-7613(1977)5<621:SOVABA>2.0.CO;2.
Scholl, M.A., Ingebritsen, S.E., 1995. Total and non-seasalt sulfate and chloride measured in bulk precipitation samples from the Kilauea volcano area, Hawaii. US Geol. Surv. Water-Resour. Invest.
PT E
D
Rep. 95–4001, 32 pp
Siegel, B.Z., Nachbar-Napai, M., Siegel, S.M., 1990. The contribution of sulfate to rainfall pH around
CE
Kilauea volcano, Hawaii. Wat. Air Soil Pollut. 52, 227–235.
Taddeucci, J., Scarlato, P., Capponi, A., Del Bello, E., Cimarelli, C., Palladino, D.M., Kueppers, U., 2012.
AC
High- speed imaging of Strombolian explosions: the ejection velocity of pyroclasts. Geophys. Res. Lett. 39, L02301.
Textor, C., Graf, H.-F., Timmreck, C., Robock, A., 2004. Emissions from volcanoes. In: Granier, C., Artaxo, P., Reeves, C. (Eds.), Emissions of Atmospheric Trace Compounds. Kluwer, Dordrecht, pp. 269–303.
Varekamp, J.C., Thomas, E., Germani, M., Buseck, P.R., 1986. Particle geochemistry of volcanic plumes
of
Etna
and
Mount
St.
Helens,
J.
Geophys.
Res.,
91,
12,233–12,248,
doi:10.1029/JB091iB12p12233.
15
ACCEPTED MANUSCRIPT
Von Glasow, R., Bobrowski, N., Kern, C., 2009. The effects of volcanic eruptions on atmospheric chemistry. Chem. Geol. 263, 131–142.
WHO, 2011. Guidelines for Drinking-water Quality. Fourth edition. WHO Press, Geneva. Switzerland, 564 pp.
PT
Zoller, W.H., Parrington, J.R., Phelan‐Kotra, J.M., 1983. Iridium enrichment in airborne particles from
AC
CE
PT E
D
MA
NU
SC
RI
Kilauea volcano: January 1983, Science, 222, 1118–1121, doi:10.1126/science.222.4628.1118.
16
ACCEPTED MANUSCRIPT Figure captions Fig. 1 Map of Stromboli island with location of sampling sites and main wind direction; Fig. 2 Rain gauge and its assemblage: a) picture of the island showing the plume moving downslope toward south-east and the FRT sampling site location; b) exploded view drawing of the rain gauge showing its different components; the perforated disks are used for supporting paper filters; c) axonometric view of the assembled rain gauge; d) front view of the rain gauge showing the container and the funnel; e) picture showing the rain sampler installed at COA station. In the centre, the orange
PT
PVC tube mounted on a metallic pole contains the rainwater sampling system illustrated in b and c; Fig. 3 Values of pH (a, b, c), TDS (d, e, f) and F concentrations (g, h, i) plotted vs distance from active
RI
vents, altitude and angle between the main propagation direction of the plume and the line
SC
connecting the vents with sampling sites. To avoid errors due to different sampling periods, plotted values are related to the temporal interval for which samples are available for all the sites (August 2015 – March 2016).
NU
Fig. 4 a) pH vs exposure time (days); b) TDS vs rain amount (mm);
Fig. 5 a) Langelier-Ludwig diagram. Samples having very similar compositions are sometimes not
MA
distinguishable in the graph because the related points could overlap; b) Temporal variation of Cl/SO4 ratio and F concentration in the COA site;
Fig. 6 Scatterplots correlating the molar abundances of cation and anion species in rainwater
D
samples: a) Cl vs SO4; b) SO4 vs Ca; c) Cl vs Na; d) Cl vs K. Black lines refer to plume’s chemical
PT E
composition, grey dotted lines represent the seawater ratio, and grey dotted-dashed lines represent the stechiometric ratio of the prevalent soluble mineralogical phases; Fig. 7 Trace elements composition range (coloured areas) of the different sampling sites (a) COA; b)
CE
FRT; d) DER). Black-dotted line represents the average composition of ash leachate (data from Bagnato et al., 2011), the blue line is the average composition of rainwater collected in roof catchment systems (Madonia et al., 2013a), and the blue-crossed diamonts are the safety limits for
AC
drinking waters established by the WHO (2011); Fig. 8 Enrichment factors (of the different sampling sites a) COA; b) FRT; c) DER) calculated for each element in rainwater, using Ti as reference element, with respect to ash composition (data from Allard et al., 2000; Bertagnini et al., 2003). Elements are ordered according to their mobility.
17
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT E
Fig. 1
18
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
Fig. 2
19
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
Fig. 3
20
AC
CE
PT E
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
Fig. 4
21
AC
CE
PT E
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
Fig. 5
22
PT E
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
Fig. 6
23
AC
CE
PT E
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
Fig. 7
24
AC
CE
PT E
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
25
AC
CE
PT E
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
Fig. 8
26
AC
CE
PT E
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
27
ACCEPTED MANUSCRIPT Table 1. Sample sites information Longitude WGS-84
Latitude WGS-84
COA DER FRT PDR SCA
15°.234189 15°.228126 15°.230482 15°.236543 15°.239173
38°.802776 38°.808216 38°.799138 38°.804936 38°.798198
Altitude (m a.s.l.) 73 25 198
Angle (°) 73 67 94
Distance from the crater (m) 2180 2070 1740
Sample type Gauge Gauge Gauge Single event Single event
AC
CE
PT E
D
MA
NU
SC
RI
PT
Site
28
ACCEPTED MANUSCRIPT
Site
Sampling date
Exposure time (day)
Rain (mm)
pH
EC
Na
K
Mg
Ca
Cl
8.03 4.51
100
12527 4.60
445 0.41
1501 0.68
447 2.13
22838 10.2
F
SO4
15 Jan 2015 15 Jan 2015
43
112
COA SCA
12 Mar 2015 12 Mar 2015
56 Single event
88
4.08 5.80
222 24.4
5.04 2.29
0.43 0.33
0.88 0.31
2.07 1.21
12.8 4.02
0.722 0.093
8.57 2.32
COA FRT
24 Apr 2015 24 Apr 2015
42 42
25 24
4.73 4.20
126 136
7.13 6.90
0.16 0.78
1.46 0.97
4.21 2.61
15.2 13.1
0.570 0.570
9.99 9.61
COA FRT PDR
21 May 2015 21 May 2015 21 May 2015
27 27 Single event
6.5 4.3
4.5 3.6 6.15
80.9
31.29 48.06 6.56
3.73 9.28 0.96
6.77 11.0 0.69
26.9 36.1 5.09
61.8 93.6 6.78
2.39 7.41 0.270
62.42 124 6.89
COA FRT
12 Aug 2015 12 Aug 2015
83 83
26 19
4.79 3.7
157
12.47 41.56
4.39 12.7
4.09 12.0
19.5 57.8
35.0 119
2.59 11.9
43.8 148
COA FRT DER
27 Jan 2016 27 Jan 2016 27 Jan 2016
168 168 168
468 333 386
3.71 3.38 4.14
70.4 120 72.4
4.49 6.15 7.34
0.00 0.97 0.72
0.77 1.47 1.12
1.71 3.89 2.52
10.2 15.3 14.3
0.505 1.11 0.378
6.91 15.9 7.88
COA FRT DER
18 Mar 2016 18 Mar 2016 18 Mar 2016
51 51 51
80 71 67
4.18 3.03 5.26
95.1 171.7 132.6
7.38 10.73 14.69
1.57 3.56 2.62
3.66 8.08 4.80
16.6 29.9 29.6
SC
RI
PT
SW COA
0.28 1.20 1.04
0.498
0.743 2.19 0.462
2981 6.16
11.56 26.98 14.06
AC
CE
PT E
D
MA
NU
Table 2. Concentration (in mg l-1) of major elements of rainwater samples collected on Stromboli Island between November 2014 and March 2016. EC is the Electrical Conductivity (S cm-1).
29
ACCEPTED MANUSCRIPT
Table 3. Concentration (g l-1) of minor and trace elements of rainwater samples collected on Stromboli Island between November 2014 and March 2016 Sampling date 15 Jan 2015
Zn 505
Ba 124
Al 210
Sb 113
Fe 33.9
Sr 18.5
Mn 11.8
B 3.12
Cu 10.0
Ni 2.45
Cr 0.400
As 0.769
Pb 0.810
Co 1.45
Rb 0.509
V 0.785
COA FRT
24 Apr 2015 24 Apr 2015
932 1317
0.363 324
222 272
66.1 140
43.2 52.9
27.9 32.9
18.7 23.8
16.4 15.5
3.31 10.3
3.80 6.65
0.853 1.58
0.817 1.13
0.299 1.01
1.16 0.881
0.233 0.856
0.206 0.583
COA FRT DER
27 Jan 2016 27 Jan 2016 27 Jan 2016
422 18.8 614
27.2 18.3 21.5
296 852 271
39.8 1.97 48.4
37.5 249 64.7
12.8 33.8 17.0
12.4 39.5 13.5
6.83 8.80 8.24
5.69 18.8 16.9
0.494 1.34 1.69
0.071 0.128 0.243
0.499 1.99 0.876
1.29 0.376 1.07
0.396 0.709 0.530
0.533 1.54 0.964
0.099 0.369 0.508
COA FRT DER
18 Mar 2016 18 Mar 2016 18 Mar 2016
143 40.7 1462
14.3 27.4 16.0
509 1410 96.5
24.5 1.69 65.9
272 996 28.7
27.5 50.8 27.9
36.3 78.5 33.5
14.5 27.7 22.0
8.19 40.3 6.52
1.81 6.54 2.23
0.136 0.334 0.150
0.612 2.42 1.01
0.521 2.02 0.183
0.810 1.45 1.25
0.575 2.02 1.24
0.299 1.75 0.286
AC
CE
PT E
D
MA
NU
SC
RI
PT
Site COA
30
ACCEPTED MANUSCRIPT Table 4: Linear correlation coefficients (r) between minor and trace element concentrations versus rainfall amount, duration of the sampling period and pH
D
pH -0.84 -0.64 -0.04 0.06 -0.77 -0.50 0.23 0.17 -0.76 -0.75 -0.70 -0.20 -0.54 0.46 -0.17 -0.71 -0.55 0.58 -0.34 -0.42 -0.82 -0.77 -0.60 -0.53 0.78
NU
SC
RI
PT
Sampling period (days) 0.05 0.02 -0.51 -0.34 -0.08 0.18 -0.80 -0.48 0.28 0.06 -0.15 0.06 -0.24 -0.49 -0.63 0.10 0.13 -0.46 -0.55 -0.41 0.19 0.21 -0.06 -0.31 -0.38
AC
CE
PT E
Al As B Ba Be Cd Co Cr Cs Cu Fe Li Mn Mo Ni Pb Rb Sb Se Sr Ti Tl U V Zn
Rain (mm) -0.02 -0.12 -0.59 -0.33 -0.16 0.00 -0.77 -0.53 0.17 0.00 -0.19 -0.03 -0.32 -0.39 -0.68 0.18 0.02 -0.39 -0.62 -0.54 0.10 0.15 -0.15 -0.31 -0.39
MA
Elements
31
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
Graphical abstract
32
ACCEPTED MANUSCRIPT Highlights Volcanic plume, sea spray aerosol and rainwater interact; Interaction occurs both in the atmosphere and after deposition;
AC
CE
PT E
D
MA
NU
SC
RI
PT
Chemistry of rain samples is conditioned by post-depositional processes.
33